Merge pull request 'fix(api): blank-name GIAS sentinel codes map to empty string, not Unknown(n)' (#27) from fix/gias-blank-name-codes into main

Reviewed-on: #27

backend/data_loader.py

@@ -4,6 +4,7 @@ Provides efficient queries with caching.
 """
 
 import logging
+import re
 
 import pandas as pd
 import numpy as np
@@ -262,25 +263,81 @@ assert "NULL AS has_sixth_form" in str(_MAIN_QUERY_NO_SIXTH_FORM), (
     "expected replacement of 's.has_sixth_form,' to have taken effect"
 )
 
+# Fallback used when marts.dim_school predates the GIAS code-dictionary
+# migration (i.e. the nightly dbt pipeline hasn't rebuilt the mart yet on
+# this DB, so it still has the old name columns instead of *_code columns).
+_MAIN_QUERY_LEGACY_NAMES = str(_MAIN_QUERY)
+_LEGACY_NAME_REPLACEMENTS = [
+    ("s.phase_code,", "s.phase,"),
+    ("s.school_type_code,", "s.school_type,"),
+    (
+        "s.religious_character_code,",
+        "s.religious_character AS religious_denomination,",
+    ),
+    ("s.status_code,", "s.status,"),
+    ("s.admissions_policy_code,", "s.admissions_policy,"),
+]
+for _old, _new in _LEGACY_NAME_REPLACEMENTS:
+    assert _old in _MAIN_QUERY_LEGACY_NAMES, (
+        f"expected {_old!r} to be present in _MAIN_QUERY before replacement"
+    )
+    _MAIN_QUERY_LEGACY_NAMES = _MAIN_QUERY_LEGACY_NAMES.replace(_old, _new)
+_MAIN_QUERY_LEGACY_NAMES = text(_MAIN_QUERY_LEGACY_NAMES)
+
+_GIAS_CODE_COLUMN_NAMES = (
+    "phase_code",
+    "school_type_code",
+    "religious_character_code",
+    "status_code",
+    "admissions_policy_code",
+)
+
+_MISSING_COLUMN_RE = re.compile(r'column "?(?:s\.)?(\w+)"? does not exist')
+
+
+def _missing_column_name(exc: Exception) -> Optional[str]:
+    """Name of the missing column from a psycopg2 UndefinedColumn error.
+
+    Inspects exc.orig (the DBAPI error), whose message names only the
+    offending column — str(exc) also embeds the full SQL statement, which
+    contains every column name and therefore must not be matched against.
+    """
+    orig = getattr(exc, "orig", None)
+    match = _MISSING_COLUMN_RE.search(str(orig) if orig is not None else str(exc))
+    return match.group(1) if match else None
+
 
 def load_school_data_as_dataframe() -> pd.DataFrame:
     """Load all school + KS2 data as a pandas DataFrame."""
     try:
         df = pd.read_sql(_MAIN_QUERY, engine)
     except sqlalchemy.exc.ProgrammingError as exc:
-        if "has_sixth_form" not in str(exc):
+        missing = _missing_column_name(exc)
+        if missing in _GIAS_CODE_COLUMN_NAMES:
+            logging.getLogger(__name__).warning(
+                "marts predate the GIAS code migration — falling back to "
+                "legacy name-column query: %s",
+                exc,
+            )
+            try:
+                df = pd.read_sql(_MAIN_QUERY_LEGACY_NAMES, engine)
+            except Exception as exc2:
+                print(f"Warning: Could not load school data from marts: {exc2}")
+                return pd.DataFrame()
+        elif missing == "has_sixth_form":
+            logging.getLogger(__name__).warning(
+                "marts.dim_school is missing has_sixth_form (pipeline hasn't "
+                "rebuilt the mart yet on this DB) — retrying without it: %s",
+                exc,
+            )
+            try:
+                df = pd.read_sql(_MAIN_QUERY_NO_SIXTH_FORM, engine)
+            except Exception as exc2:
+                print(f"Warning: Could not load school data from marts: {exc2}")
+                return pd.DataFrame()
+        else:
             print(f"Warning: Could not load school data from marts: {exc}")
             return pd.DataFrame()
-        logging.getLogger(__name__).warning(
-            "marts.dim_school is missing has_sixth_form (pipeline hasn't "
-            "rebuilt the mart yet on this DB) — retrying without it: %s",
-            exc,
-        )
-        try:
-            df = pd.read_sql(_MAIN_QUERY_NO_SIXTH_FORM, engine)
-        except Exception as exc2:
-            print(f"Warning: Could not load school data from marts: {exc2}")
-            return pd.DataFrame()
     except Exception as exc:
         print(f"Warning: Could not load school data from marts: {exc}")
         return pd.DataFrame()

backend/gias_codes.py

@@ -78,6 +78,7 @@ OFFICIAL_SIXTH_FORM: dict[int, str] = {
     0: "Not applicable",
     1: "Has a sixth form",
     2: "Does not have a sixth form",
+    9: "",
 }
 
 RELIGIOUS_CHARACTER: dict[int, str] = {
@@ -128,12 +129,14 @@ RELIGIOUS_CHARACTER: dict[int, str] = {
     47: "Reformed Baptist",
     48: "Roman Catholic/Anglican",
     49: "Sunni Deobandi",
+    99: "",
 }
 
 ADMISSIONS_POLICY: dict[int, str] = {
     0: "Not applicable",
     2: "Selective",
     4: "Non-selective",
+    9: "",
 }
 
 

backend/tests/test_gias_codes.py

@@ -81,3 +81,15 @@ def test_seed_matches_dictionaries():
         for row in csv.DictReader(fh):
             seed[row["field"]][int(row["code"])] = row["name"]
     assert seed == fields
+
+
+def test_blank_name_sentinel_codes_map_to_empty_string():
+    """GIAS carries codes whose (name) column is blank — e.g. ReligiousCharacter
+    99 (~4k schools) and AdmissionsPolicy 9 (~5.6k schools). The old name
+    pipeline served these as empty strings; the dictionaries must reproduce
+    that ("" is falsy, so UI tag heuristics stay silent) rather than letting
+    them hit the "Unknown (<code>)" path meant for genuinely new codes."""
+    assert RELIGIOUS_CHARACTER[99] == ""
+    assert ADMISSIONS_POLICY[9] == ""
+    assert translate(99, RELIGIOUS_CHARACTER) == ""
+    assert translate(9, ADMISSIONS_POLICY) == ""

backend/tests/test_gias_translation.py

@@ -4,7 +4,7 @@ rest of the backend sees must carry today's name strings."""
 import numpy as np
 import pandas as pd
 
-from backend.data_loader import translate_gias_code_columns
+from backend.data_loader import _missing_column_name, translate_gias_code_columns
 from backend.gias_codes import ESTABLISHMENT_STATUS, PHASE_OF_EDUCATION
 
 
@@ -42,3 +42,91 @@ def test_missing_code_columns_are_a_noop():
     out = translate_gias_code_columns(df)
     assert out.iloc[0]["phase"] == "Primary"
     assert out.iloc[0]["status"] == "Open"
+
+
+def _fake_exc(orig_message):
+    """A stand-in for sqlalchemy.exc.ProgrammingError: str(exc) embeds the
+    full SQL statement (deliberately containing every column name below, to
+    prove the matcher doesn't fall back to it), while .orig carries the real
+    DBAPI error message naming only the offending column."""
+    exc = Exception(
+        "SELECT s.phase_code, s.school_type_code, s.religious_character_code, "
+        "s.status_code, s.admissions_policy_code, s.has_sixth_form FROM ... "
+        f"[SQL: ...] (Background on this error at: https://...)"
+    )
+    exc.orig = Exception(orig_message) if orig_message is not None else None
+    return exc
+
+
+def test_missing_column_name_quoted():
+    assert _missing_column_name(_fake_exc('column "phase_code" does not exist')) == "phase_code"
+
+
+def test_missing_column_name_unquoted():
+    assert _missing_column_name(_fake_exc("column phase_code does not exist")) == "phase_code"
+
+
+def test_missing_column_name_table_prefixed():
+    assert (
+        _missing_column_name(_fake_exc("column s.has_sixth_form does not exist"))
+        == "has_sixth_form"
+    )
+
+
+def test_missing_column_name_no_match_returns_none():
+    assert _missing_column_name(_fake_exc("relation \"marts.dim_school\" does not exist")) is None
+
+
+def test_load_school_data_survives_premigration_marts(monkeypatch):
+    """Real prod state until the nightly pipeline first rebuilds the mart with
+    the GIAS code columns: marts.dim_school still has the old name columns
+    (phase, school_type, religious_character, status, admissions_policy)
+    instead of the new *_code columns. The first query raises UndefinedColumn
+    on s.phase_code; load_school_data_as_dataframe must retry with the
+    legacy name-column query rather than swallow the error and return (and
+    then have load_school_data cache) an empty DataFrame."""
+    import sqlalchemy.exc
+    from backend import data_loader
+
+    data_loader._df_cache = None
+    data_loader._df_latest_cache = None
+
+    good_df = pd.DataFrame(
+        [
+            {
+                "urn": 1,
+                "school_name": "Legacy School",
+                "phase": "Primary",
+                "school_type": "Academy",
+                "status": "Open",
+            }
+        ]
+    )
+    calls = []
+
+    def fake_read_sql(query, con):
+        calls.append(query)
+        if len(calls) == 1:
+            raise sqlalchemy.exc.ProgrammingError(
+                statement=str(data_loader._MAIN_QUERY),
+                params=None,
+                orig=Exception(
+                    "(psycopg2.errors.UndefinedColumn) column s.phase_code "
+                    "does not exist\nLINE 5:         s.phase_code,"
+                ),
+            )
+        return good_df.copy()
+
+    monkeypatch.setattr(data_loader.pd, "read_sql", fake_read_sql)
+
+    try:
+        df = data_loader.load_school_data_as_dataframe()
+    finally:
+        data_loader._df_cache = None
+        data_loader._df_latest_cache = None
+
+    assert len(calls) == 2, "must retry with the legacy name-column query variant"
+    assert calls[1] is data_loader._MAIN_QUERY_LEGACY_NAMES
+    assert not df.empty
+    assert df["phase"].iloc[0] == "Primary"
+    assert df["status"].iloc[0] == "Open"

backend/tests/test_sixth_form_flag.py

@@ -148,10 +148,14 @@ def test_load_school_data_survives_missing_has_sixth_form_column(monkeypatch):
     def fake_read_sql(query, con):
         calls.append(query)
         if len(calls) == 1:
+            # The statement text still contains phase_code, school_type_code,
+            # etc. (it's the full _MAIN_QUERY SELECT list) — that's exactly
+            # the collision this test guards against: matching must be done
+            # against exc.orig (the DBAPI error), not str(exc)/the statement.
             raise sqlalchemy.exc.ProgrammingError(
-                "SELECT ...",
-                None,
-                Exception(
+                statement=str(data_loader._MAIN_QUERY),
+                params=None,
+                orig=Exception(
                     "(psycopg2.errors.UndefinedColumn) column s.has_sixth_form "
                     "does not exist"
                 ),

pipeline/scripts/generate_gias_codes.py

@@ -94,13 +94,23 @@ def main() -> None:
     for code_col, name_col, dict_name, field_key in FIELDS:
         pairs = (
             df[[code_col, name_col]]
-            .loc[lambda d: (d[code_col] != "") & (d[name_col] != "")]
+            .loc[lambda d: d[code_col] != ""]
             .drop_duplicates()
         )
-        mapping = sorted((int(c), n) for c, n in pairs.itertuples(index=False))
-        dupes = len(mapping) - len({c for c, _ in mapping})
-        if dupes:
-            sys.exit(f"{code_col}: {dupes} codes map to multiple names — investigate before generating")
+        by_code: dict[int, set] = {}
+        for c, n in pairs.itertuples(index=False):
+            by_code.setdefault(int(c), set()).add(n)
+        mapping = []
+        for code, names in sorted(by_code.items()):
+            named = sorted(n for n in names if n != "")
+            if len(named) > 1:
+                sys.exit(f"{code_col}: code {code} maps to multiple names {named} — investigate before generating")
+            # Codes that only ever appear with a blank (name) are GIAS
+            # "not recorded" sentinels (e.g. ReligiousCharacter 99,
+            # AdmissionsPolicy 9). Map them to "" so the API serves the same
+            # empty string the old name pipeline did — the "Unknown (<code>)"
+            # path is reserved for genuinely new codes.
+            mapping.append((code, named[0] if named else ""))
         lines = [f"{dict_name}: dict[int, str] = {{"]
         for code, name in mapping:
             escaped = name.replace('"', '\\"')

pipeline/scripts/gias_codes.py

@@ -78,6 +78,7 @@ OFFICIAL_SIXTH_FORM: dict[int, str] = {
     0: "Not applicable",
     1: "Has a sixth form",
     2: "Does not have a sixth form",
+    9: "",
 }
 
 RELIGIOUS_CHARACTER: dict[int, str] = {
@@ -128,12 +129,14 @@ RELIGIOUS_CHARACTER: dict[int, str] = {
     47: "Reformed Baptist",
     48: "Roman Catholic/Anglican",
     49: "Sunni Deobandi",
+    99: "",
 }
 
 ADMISSIONS_POLICY: dict[int, str] = {
     0: "Not applicable",
     2: "Selective",
     4: "Non-selective",
+    9: "",
 }
 
 

pipeline/transform/models/marts/_marts_schema.yml

@@ -42,12 +42,12 @@ models:
         tests:
           - accepted_values:
               severity: warn
-              values: [0, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 24, 25, 26, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49]
+              values: [0, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 24, 25, 26, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 99]
       - name: admissions_policy_code
         tests:
           - accepted_values:
               severity: warn
-              values: [0, 2, 4]
+              values: [0, 2, 4, 9]
 
   - name: dim_location
     description: School location dimension with PostGIS geometry

pipeline/transform/seeds/gias_code_names.csv

@@ -53,6 +53,7 @@ phase_of_education,7,All-through
 official_sixth_form,0,Not applicable
 official_sixth_form,1,Has a sixth form
 official_sixth_form,2,Does not have a sixth form
+official_sixth_form,9,
 religious_character,0,Does not apply
 religious_character,2,Church of England
 religious_character,3,Roman Catholic
@@ -100,6 +101,8 @@ religious_character,46,Protestant/Evangelical
 religious_character,47,Reformed Baptist
 religious_character,48,Roman Catholic/Anglican
 religious_character,49,Sunni Deobandi
+religious_character,99,
 admissions_policy,0,Not applicable
 admissions_policy,2,Selective
 admissions_policy,4,Non-selective
+admissions_policy,9,